Hugging Face's logo

AbstractFramework
/
qwen-image-4bit

Text-to-Image
Safetensors
MLX
mlx-gen
mflux
apple-silicon
4-bit precision
mixed-q4
mixed-q4-q8
qwen
qwen-image
Model card Files
xet
 Community
lpalbou commited on
Commit
70bc8a6
·
verified ·
1 Parent(s): 7dc393d

Upload folder using huggingface_hub

Browse files
Files changed (17) hide show
  1. .gitattributes +1 -0
  2. README.md +66 -0
  3. text_encoder/0.safetensors +3 -0
  4. text_encoder/1.safetensors +3 -0
  5. text_encoder/model.safetensors.index.json +769 -0
  6. tokenizer/tokenizer.json +3 -0
  7. tokenizer/tokenizer_config.json +10 -0
  8. transformer/0.safetensors +3 -0
  9. transformer/1.safetensors +3 -0
  10. transformer/2.safetensors +3 -0
  11. transformer/3.safetensors +3 -0
  12. transformer/4.safetensors +3 -0
  13. transformer/5.safetensors +3 -0
  14. transformer/6.safetensors +3 -0
  15. transformer/model.safetensors.index.json +0 -0
  16. vae/0.safetensors +3 -0
  17. vae/model.safetensors.index.json +200 -0
.gitattributes CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
 
 
33
  *.zip filter=lfs diff=lfs merge=lfs -text
34
  *.zst filter=lfs diff=lfs merge=lfs -text
35
  *tfevents* filter=lfs diff=lfs merge=lfs -text
36
+ tokenizer/tokenizer.json filter=lfs diff=lfs merge=lfs -text
README.md ADDED
@@ -0,0 +1,66 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ ---
2
+ license: apache-2.0
3
+ base_model: Qwen/Qwen-Image
4
+ pipeline_tag: text-to-image
5
+ library_name: mlx-gen
6
+ tags:
7
+ - mlx
8
+ - mlx-gen
9
+ - mflux
10
+ - apple-silicon
11
+ - 4-bit
12
+ - mixed-q4
13
+ - mixed-q4-q8
14
+ - qwen
15
+ - qwen-image
16
+ ---
17
+ # qwen-image-4bit
18
+
19
+ This repository contains MLX-Gen saved weights for `Qwen/Qwen-Image`. The checkpoint is designed for local Apple Silicon inference with [`mlx-gen`](https://github.com/lpalbou/mlx-gen).
20
+
21
+ It uses the mflux/MLX saved-weight layout and MLX quantization tensors. It is not a Diffusers or Transformers `from_pretrained()` checkpoint.
22
+
23
+ ## Source Model
24
+
25
+ Original model: [`Qwen/Qwen-Image`](https://huggingface.co/Qwen/Qwen-Image).
26
+
27
+ ## Quantization
28
+
29
+ This is a mixed q4/q8 checkpoint for Qwen Image generation and editing. Fully q4 Qwen checkpoints can lose coherent generative behavior, so MLX-Gen uses a mixed policy:
30
+
31
+ - q4 for most Qwen transformer attention, feed-forward, and projection linears.
32
+ - q8 for Qwen `*.img_mod_linear` transformer modulation layers.
33
+ - q4 for group64-compatible Qwen text-encoder language linears.
34
+ - q8 for group64-compatible Qwen text-encoder visual linears.
35
+ - BF16 for the VAE, norms, embeddings, and linears that are not MLX group64-compatible.
36
+
37
+ See the [MLX-Gen quantization docs](https://github.com/lpalbou/mlx-gen/blob/main/docs/quantization.md) for the current mixed q4/q8 policy and compatibility notes.
38
+
39
+ ## Compatibility
40
+
41
+ Requires `mlx-gen >= 0.18.2`.
42
+
43
+ Generated with `mlx-gen 0.18.2`.
44
+
45
+ Use the `mlxgen` command and Python import path for new MLX-Gen projects.
46
+
47
+ ## Usage
48
+
49
+ ```bash
50
+ python -m pip install -U mlx-gen
51
+
52
+ mlxgen download --model AbstractFramework/qwen-image-4bit
53
+
54
+ mlxgen generate \
55
+ --model AbstractFramework/qwen-image-4bit \
56
+ --prompt "Your prompt here" \
57
+ --steps 20 \
58
+ --seed 42 \
59
+ --output image.png
60
+ ```
61
+
62
+ ## Attribution
63
+
64
+ MLX-Gen is based on [mflux](https://github.com/filipstrand/mflux) by Filip Strand and the original mflux contributors. This model card is generated by MLX-Gen so derived checkpoints keep that attribution visible.
65
+
66
+ Quantized and contributed by [@lpalbou](https://huggingface.co/lpalbou).
text_encoder/0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b17f5c5e9a67d4e1e99661429afe277cacc6dfa1b29d505bf38d8b5f1e4ad0da
3
+ size 2142182659
text_encoder/1.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:c379ed1c78714f3aacbd5a6e3ad708f0cc33c4a4f6689ef3010e294dbfd93d54
3
+ size 1835614566
text_encoder/model.safetensors.index.json ADDED
@@ -0,0 +1,769 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "quantization_level": "4",
4
+ "mflux_version": "0.18.2"
5
+ },
6
+ "weight_map": {
7
+ "encoder.embed_tokens.weight": "0.safetensors",
8
+ "encoder.embed_tokens.scales": "0.safetensors",
9
+ "encoder.embed_tokens.biases": "0.safetensors",
10
+ "encoder.layers.0.input_layernorm.weight": "0.safetensors",
11
+ "encoder.layers.0.self_attn.q_proj.weight": "0.safetensors",
12
+ "encoder.layers.0.self_attn.q_proj.scales": "0.safetensors",
13
+ "encoder.layers.0.self_attn.q_proj.biases": "0.safetensors",
14
+ "encoder.layers.0.self_attn.q_proj.bias": "0.safetensors",
15
+ "encoder.layers.0.self_attn.k_proj.weight": "0.safetensors",
16
+ "encoder.layers.0.self_attn.k_proj.scales": "0.safetensors",
17
+ "encoder.layers.0.self_attn.k_proj.biases": "0.safetensors",
18
+ "encoder.layers.0.self_attn.k_proj.bias": "0.safetensors",
19
+ "encoder.layers.0.self_attn.v_proj.weight": "0.safetensors",
20
+ "encoder.layers.0.self_attn.v_proj.scales": "0.safetensors",
21
+ "encoder.layers.0.self_attn.v_proj.biases": "0.safetensors",
22
+ "encoder.layers.0.self_attn.v_proj.bias": "0.safetensors",
23
+ "encoder.layers.0.self_attn.o_proj.weight": "0.safetensors",
24
+ "encoder.layers.0.self_attn.o_proj.scales": "0.safetensors",
25
+ "encoder.layers.0.self_attn.o_proj.biases": "0.safetensors",
26
+ "encoder.layers.0.self_attn.rotary_emb.inv_freq": "0.safetensors",
27
+ "encoder.layers.0.post_attention_layernorm.weight": "0.safetensors",
28
+ "encoder.layers.0.mlp.gate_proj.weight": "0.safetensors",
29
+ "encoder.layers.0.mlp.gate_proj.scales": "0.safetensors",
30
+ "encoder.layers.0.mlp.gate_proj.biases": "0.safetensors",
31
+ "encoder.layers.0.mlp.up_proj.weight": "0.safetensors",
32
+ "encoder.layers.0.mlp.up_proj.scales": "0.safetensors",
33
+ "encoder.layers.0.mlp.up_proj.biases": "0.safetensors",
34
+ "encoder.layers.0.mlp.down_proj.weight": "0.safetensors",
35
+ "encoder.layers.0.mlp.down_proj.scales": "0.safetensors",
36
+ "encoder.layers.0.mlp.down_proj.biases": "0.safetensors",
37
+ "encoder.layers.1.input_layernorm.weight": "0.safetensors",
38
+ "encoder.layers.1.self_attn.q_proj.weight": "0.safetensors",
39
+ "encoder.layers.1.self_attn.q_proj.scales": "0.safetensors",
40
+ "encoder.layers.1.self_attn.q_proj.biases": "0.safetensors",
41
+ "encoder.layers.1.self_attn.q_proj.bias": "0.safetensors",
42
+ "encoder.layers.1.self_attn.k_proj.weight": "0.safetensors",
43
+ "encoder.layers.1.self_attn.k_proj.scales": "0.safetensors",
44
+ "encoder.layers.1.self_attn.k_proj.biases": "0.safetensors",
45
+ "encoder.layers.1.self_attn.k_proj.bias": "0.safetensors",
46
+ "encoder.layers.1.self_attn.v_proj.weight": "0.safetensors",
47
+ "encoder.layers.1.self_attn.v_proj.scales": "0.safetensors",
48
+ "encoder.layers.1.self_attn.v_proj.biases": "0.safetensors",
49
+ "encoder.layers.1.self_attn.v_proj.bias": "0.safetensors",
50
+ "encoder.layers.1.self_attn.o_proj.weight": "0.safetensors",
51
+ "encoder.layers.1.self_attn.o_proj.scales": "0.safetensors",
52
+ "encoder.layers.1.self_attn.o_proj.biases": "0.safetensors",
53
+ "encoder.layers.1.self_attn.rotary_emb.inv_freq": "0.safetensors",
54
+ "encoder.layers.1.post_attention_layernorm.weight": "0.safetensors",
55
+ "encoder.layers.1.mlp.gate_proj.weight": "0.safetensors",
56
+ "encoder.layers.1.mlp.gate_proj.scales": "0.safetensors",
57
+ "encoder.layers.1.mlp.gate_proj.biases": "0.safetensors",
58
+ "encoder.layers.1.mlp.up_proj.weight": "0.safetensors",
59
+ "encoder.layers.1.mlp.up_proj.scales": "0.safetensors",
60
+ "encoder.layers.1.mlp.up_proj.biases": "0.safetensors",
61
+ "encoder.layers.1.mlp.down_proj.weight": "0.safetensors",
62
+ "encoder.layers.1.mlp.down_proj.scales": "0.safetensors",
63
+ "encoder.layers.1.mlp.down_proj.biases": "0.safetensors",
64
+ "encoder.layers.2.input_layernorm.weight": "0.safetensors",
65
+ "encoder.layers.2.self_attn.q_proj.weight": "0.safetensors",
66
+ "encoder.layers.2.self_attn.q_proj.scales": "0.safetensors",
67
+ "encoder.layers.2.self_attn.q_proj.biases": "0.safetensors",
68
+ "encoder.layers.2.self_attn.q_proj.bias": "0.safetensors",
69
+ "encoder.layers.2.self_attn.k_proj.weight": "0.safetensors",
70
+ "encoder.layers.2.self_attn.k_proj.scales": "0.safetensors",
71
+ "encoder.layers.2.self_attn.k_proj.biases": "0.safetensors",
72
+ "encoder.layers.2.self_attn.k_proj.bias": "0.safetensors",
73
+ "encoder.layers.2.self_attn.v_proj.weight": "0.safetensors",
74
+ "encoder.layers.2.self_attn.v_proj.scales": "0.safetensors",
75
+ "encoder.layers.2.self_attn.v_proj.biases": "0.safetensors",
76
+ "encoder.layers.2.self_attn.v_proj.bias": "0.safetensors",
77
+ "encoder.layers.2.self_attn.o_proj.weight": "0.safetensors",
78
+ "encoder.layers.2.self_attn.o_proj.scales": "0.safetensors",
79
+ "encoder.layers.2.self_attn.o_proj.biases": "0.safetensors",
80
+ "encoder.layers.2.self_attn.rotary_emb.inv_freq": "0.safetensors",
81
+ "encoder.layers.2.post_attention_layernorm.weight": "0.safetensors",
82
+ "encoder.layers.2.mlp.gate_proj.weight": "0.safetensors",
83
+ "encoder.layers.2.mlp.gate_proj.scales": "0.safetensors",
84
+ "encoder.layers.2.mlp.gate_proj.biases": "0.safetensors",
85
+ "encoder.layers.2.mlp.up_proj.weight": "0.safetensors",
86
+ "encoder.layers.2.mlp.up_proj.scales": "0.safetensors",
87
+ "encoder.layers.2.mlp.up_proj.biases": "0.safetensors",
88
+ "encoder.layers.2.mlp.down_proj.weight": "0.safetensors",
89
+ "encoder.layers.2.mlp.down_proj.scales": "0.safetensors",
90
+ "encoder.layers.2.mlp.down_proj.biases": "0.safetensors",
91
+ "encoder.layers.3.input_layernorm.weight": "0.safetensors",
92
+ "encoder.layers.3.self_attn.q_proj.weight": "0.safetensors",
93
+ "encoder.layers.3.self_attn.q_proj.scales": "0.safetensors",
94
+ "encoder.layers.3.self_attn.q_proj.biases": "0.safetensors",
95
+ "encoder.layers.3.self_attn.q_proj.bias": "0.safetensors",
96
+ "encoder.layers.3.self_attn.k_proj.weight": "0.safetensors",
97
+ "encoder.layers.3.self_attn.k_proj.scales": "0.safetensors",
98
+ "encoder.layers.3.self_attn.k_proj.biases": "0.safetensors",
99
+ "encoder.layers.3.self_attn.k_proj.bias": "0.safetensors",
100
+ "encoder.layers.3.self_attn.v_proj.weight": "0.safetensors",
101
+ "encoder.layers.3.self_attn.v_proj.scales": "0.safetensors",
102
+ "encoder.layers.3.self_attn.v_proj.biases": "0.safetensors",
103
+ "encoder.layers.3.self_attn.v_proj.bias": "0.safetensors",
104
+ "encoder.layers.3.self_attn.o_proj.weight": "0.safetensors",
105
+ "encoder.layers.3.self_attn.o_proj.scales": "0.safetensors",
106
+ "encoder.layers.3.self_attn.o_proj.biases": "0.safetensors",
107
+ "encoder.layers.3.self_attn.rotary_emb.inv_freq": "0.safetensors",
108
+ "encoder.layers.3.post_attention_layernorm.weight": "0.safetensors",
109
+ "encoder.layers.3.mlp.gate_proj.weight": "0.safetensors",
110
+ "encoder.layers.3.mlp.gate_proj.scales": "0.safetensors",
111
+ "encoder.layers.3.mlp.gate_proj.biases": "0.safetensors",
112
+ "encoder.layers.3.mlp.up_proj.weight": "0.safetensors",
113
+ "encoder.layers.3.mlp.up_proj.scales": "0.safetensors",
114
+ "encoder.layers.3.mlp.up_proj.biases": "0.safetensors",
115
+ "encoder.layers.3.mlp.down_proj.weight": "0.safetensors",
116
+ "encoder.layers.3.mlp.down_proj.scales": "0.safetensors",
117
+ "encoder.layers.3.mlp.down_proj.biases": "0.safetensors",
118
+ "encoder.layers.4.input_layernorm.weight": "0.safetensors",
119
+ "encoder.layers.4.self_attn.q_proj.weight": "0.safetensors",
120
+ "encoder.layers.4.self_attn.q_proj.scales": "0.safetensors",
121
+ "encoder.layers.4.self_attn.q_proj.biases": "0.safetensors",
122
+ "encoder.layers.4.self_attn.q_proj.bias": "0.safetensors",
123
+ "encoder.layers.4.self_attn.k_proj.weight": "0.safetensors",
124
+ "encoder.layers.4.self_attn.k_proj.scales": "0.safetensors",
125
+ "encoder.layers.4.self_attn.k_proj.biases": "0.safetensors",
126
+ "encoder.layers.4.self_attn.k_proj.bias": "0.safetensors",
127
+ "encoder.layers.4.self_attn.v_proj.weight": "0.safetensors",
128
+ "encoder.layers.4.self_attn.v_proj.scales": "0.safetensors",
129
+ "encoder.layers.4.self_attn.v_proj.biases": "0.safetensors",
130
+ "encoder.layers.4.self_attn.v_proj.bias": "0.safetensors",
131
+ "encoder.layers.4.self_attn.o_proj.weight": "0.safetensors",
132
+ "encoder.layers.4.self_attn.o_proj.scales": "0.safetensors",
133
+ "encoder.layers.4.self_attn.o_proj.biases": "0.safetensors",
134
+ "encoder.layers.4.self_attn.rotary_emb.inv_freq": "0.safetensors",
135
+ "encoder.layers.4.post_attention_layernorm.weight": "0.safetensors",
136
+ "encoder.layers.4.mlp.gate_proj.weight": "0.safetensors",
137
+ "encoder.layers.4.mlp.gate_proj.scales": "0.safetensors",
138
+ "encoder.layers.4.mlp.gate_proj.biases": "0.safetensors",
139
+ "encoder.layers.4.mlp.up_proj.weight": "0.safetensors",
140
+ "encoder.layers.4.mlp.up_proj.scales": "0.safetensors",
141
+ "encoder.layers.4.mlp.up_proj.biases": "0.safetensors",
142
+ "encoder.layers.4.mlp.down_proj.weight": "0.safetensors",
143
+ "encoder.layers.4.mlp.down_proj.scales": "0.safetensors",
144
+ "encoder.layers.4.mlp.down_proj.biases": "0.safetensors",
145
+ "encoder.layers.5.input_layernorm.weight": "0.safetensors",
146
+ "encoder.layers.5.self_attn.q_proj.weight": "0.safetensors",
147
+ "encoder.layers.5.self_attn.q_proj.scales": "0.safetensors",
148
+ "encoder.layers.5.self_attn.q_proj.biases": "0.safetensors",
149
+ "encoder.layers.5.self_attn.q_proj.bias": "0.safetensors",
150
+ "encoder.layers.5.self_attn.k_proj.weight": "0.safetensors",
151
+ "encoder.layers.5.self_attn.k_proj.scales": "0.safetensors",
152
+ "encoder.layers.5.self_attn.k_proj.biases": "0.safetensors",
153
+ "encoder.layers.5.self_attn.k_proj.bias": "0.safetensors",
154
+ "encoder.layers.5.self_attn.v_proj.weight": "0.safetensors",
155
+ "encoder.layers.5.self_attn.v_proj.scales": "0.safetensors",
156
+ "encoder.layers.5.self_attn.v_proj.biases": "0.safetensors",
157
+ "encoder.layers.5.self_attn.v_proj.bias": "0.safetensors",
158
+ "encoder.layers.5.self_attn.o_proj.weight": "0.safetensors",
159
+ "encoder.layers.5.self_attn.o_proj.scales": "0.safetensors",
160
+ "encoder.layers.5.self_attn.o_proj.biases": "0.safetensors",
161
+ "encoder.layers.5.self_attn.rotary_emb.inv_freq": "0.safetensors",
162
+ "encoder.layers.5.post_attention_layernorm.weight": "0.safetensors",
163
+ "encoder.layers.5.mlp.gate_proj.weight": "0.safetensors",
164
+ "encoder.layers.5.mlp.gate_proj.scales": "0.safetensors",
165
+ "encoder.layers.5.mlp.gate_proj.biases": "0.safetensors",
166
+ "encoder.layers.5.mlp.up_proj.weight": "0.safetensors",
167
+ "encoder.layers.5.mlp.up_proj.scales": "0.safetensors",
168
+ "encoder.layers.5.mlp.up_proj.biases": "0.safetensors",
169
+ "encoder.layers.5.mlp.down_proj.weight": "0.safetensors",
170
+ "encoder.layers.5.mlp.down_proj.scales": "0.safetensors",
171
+ "encoder.layers.5.mlp.down_proj.biases": "0.safetensors",
172
+ "encoder.layers.6.input_layernorm.weight": "0.safetensors",
173
+ "encoder.layers.6.self_attn.q_proj.weight": "0.safetensors",
174
+ "encoder.layers.6.self_attn.q_proj.scales": "0.safetensors",
175
+ "encoder.layers.6.self_attn.q_proj.biases": "0.safetensors",
176
+ "encoder.layers.6.self_attn.q_proj.bias": "0.safetensors",
177
+ "encoder.layers.6.self_attn.k_proj.weight": "0.safetensors",
178
+ "encoder.layers.6.self_attn.k_proj.scales": "0.safetensors",
179
+ "encoder.layers.6.self_attn.k_proj.biases": "0.safetensors",
180
+ "encoder.layers.6.self_attn.k_proj.bias": "0.safetensors",
181
+ "encoder.layers.6.self_attn.v_proj.weight": "0.safetensors",
182
+ "encoder.layers.6.self_attn.v_proj.scales": "0.safetensors",
183
+ "encoder.layers.6.self_attn.v_proj.biases": "0.safetensors",
184
+ "encoder.layers.6.self_attn.v_proj.bias": "0.safetensors",
185
+ "encoder.layers.6.self_attn.o_proj.weight": "0.safetensors",
186
+ "encoder.layers.6.self_attn.o_proj.scales": "0.safetensors",
187
+ "encoder.layers.6.self_attn.o_proj.biases": "0.safetensors",
188
+ "encoder.layers.6.self_attn.rotary_emb.inv_freq": "0.safetensors",
189
+ "encoder.layers.6.post_attention_layernorm.weight": "0.safetensors",
190
+ "encoder.layers.6.mlp.gate_proj.weight": "0.safetensors",
191
+ "encoder.layers.6.mlp.gate_proj.scales": "0.safetensors",
192
+ "encoder.layers.6.mlp.gate_proj.biases": "0.safetensors",
193
+ "encoder.layers.6.mlp.up_proj.weight": "0.safetensors",
194
+ "encoder.layers.6.mlp.up_proj.scales": "0.safetensors",
195
+ "encoder.layers.6.mlp.up_proj.biases": "0.safetensors",
196
+ "encoder.layers.6.mlp.down_proj.weight": "0.safetensors",
197
+ "encoder.layers.6.mlp.down_proj.scales": "0.safetensors",
198
+ "encoder.layers.6.mlp.down_proj.biases": "0.safetensors",
199
+ "encoder.layers.7.input_layernorm.weight": "0.safetensors",
200
+ "encoder.layers.7.self_attn.q_proj.weight": "0.safetensors",
201
+ "encoder.layers.7.self_attn.q_proj.scales": "0.safetensors",
202
+ "encoder.layers.7.self_attn.q_proj.biases": "0.safetensors",
203
+ "encoder.layers.7.self_attn.q_proj.bias": "0.safetensors",
204
+ "encoder.layers.7.self_attn.k_proj.weight": "0.safetensors",
205
+ "encoder.layers.7.self_attn.k_proj.scales": "0.safetensors",
206
+ "encoder.layers.7.self_attn.k_proj.biases": "0.safetensors",
207
+ "encoder.layers.7.self_attn.k_proj.bias": "0.safetensors",
208
+ "encoder.layers.7.self_attn.v_proj.weight": "0.safetensors",
209
+ "encoder.layers.7.self_attn.v_proj.scales": "0.safetensors",
210
+ "encoder.layers.7.self_attn.v_proj.biases": "0.safetensors",
211
+ "encoder.layers.7.self_attn.v_proj.bias": "0.safetensors",
212
+ "encoder.layers.7.self_attn.o_proj.weight": "0.safetensors",
213
+ "encoder.layers.7.self_attn.o_proj.scales": "0.safetensors",
214
+ "encoder.layers.7.self_attn.o_proj.biases": "0.safetensors",
215
+ "encoder.layers.7.self_attn.rotary_emb.inv_freq": "0.safetensors",
216
+ "encoder.layers.7.post_attention_layernorm.weight": "0.safetensors",
217
+ "encoder.layers.7.mlp.gate_proj.weight": "0.safetensors",
218
+ "encoder.layers.7.mlp.gate_proj.scales": "0.safetensors",
219
+ "encoder.layers.7.mlp.gate_proj.biases": "0.safetensors",
220
+ "encoder.layers.7.mlp.up_proj.weight": "0.safetensors",
221
+ "encoder.layers.7.mlp.up_proj.scales": "0.safetensors",
222
+ "encoder.layers.7.mlp.up_proj.biases": "0.safetensors",
223
+ "encoder.layers.7.mlp.down_proj.weight": "0.safetensors",
224
+ "encoder.layers.7.mlp.down_proj.scales": "0.safetensors",
225
+ "encoder.layers.7.mlp.down_proj.biases": "0.safetensors",
226
+ "encoder.layers.8.input_layernorm.weight": "0.safetensors",
227
+ "encoder.layers.8.self_attn.q_proj.weight": "0.safetensors",
228
+ "encoder.layers.8.self_attn.q_proj.scales": "0.safetensors",
229
+ "encoder.layers.8.self_attn.q_proj.biases": "0.safetensors",
230
+ "encoder.layers.8.self_attn.q_proj.bias": "0.safetensors",
231
+ "encoder.layers.8.self_attn.k_proj.weight": "0.safetensors",
232
+ "encoder.layers.8.self_attn.k_proj.scales": "0.safetensors",
233
+ "encoder.layers.8.self_attn.k_proj.biases": "0.safetensors",
234
+ "encoder.layers.8.self_attn.k_proj.bias": "0.safetensors",
235
+ "encoder.layers.8.self_attn.v_proj.weight": "0.safetensors",
236
+ "encoder.layers.8.self_attn.v_proj.scales": "0.safetensors",
237
+ "encoder.layers.8.self_attn.v_proj.biases": "0.safetensors",
238
+ "encoder.layers.8.self_attn.v_proj.bias": "0.safetensors",
239
+ "encoder.layers.8.self_attn.o_proj.weight": "0.safetensors",
240
+ "encoder.layers.8.self_attn.o_proj.scales": "0.safetensors",
241
+ "encoder.layers.8.self_attn.o_proj.biases": "0.safetensors",
242
+ "encoder.layers.8.self_attn.rotary_emb.inv_freq": "0.safetensors",
243
+ "encoder.layers.8.post_attention_layernorm.weight": "0.safetensors",
244
+ "encoder.layers.8.mlp.gate_proj.weight": "0.safetensors",
245
+ "encoder.layers.8.mlp.gate_proj.scales": "0.safetensors",
246
+ "encoder.layers.8.mlp.gate_proj.biases": "0.safetensors",
247
+ "encoder.layers.8.mlp.up_proj.weight": "0.safetensors",
248
+ "encoder.layers.8.mlp.up_proj.scales": "0.safetensors",
249
+ "encoder.layers.8.mlp.up_proj.biases": "0.safetensors",
250
+ "encoder.layers.8.mlp.down_proj.weight": "0.safetensors",
251
+ "encoder.layers.8.mlp.down_proj.scales": "0.safetensors",
252
+ "encoder.layers.8.mlp.down_proj.biases": "0.safetensors",
253
+ "encoder.layers.9.input_layernorm.weight": "0.safetensors",
254
+ "encoder.layers.9.self_attn.q_proj.weight": "0.safetensors",
255
+ "encoder.layers.9.self_attn.q_proj.scales": "0.safetensors",
256
+ "encoder.layers.9.self_attn.q_proj.biases": "0.safetensors",
257
+ "encoder.layers.9.self_attn.q_proj.bias": "0.safetensors",
258
+ "encoder.layers.9.self_attn.k_proj.weight": "0.safetensors",
259
+ "encoder.layers.9.self_attn.k_proj.scales": "0.safetensors",
260
+ "encoder.layers.9.self_attn.k_proj.biases": "0.safetensors",
261
+ "encoder.layers.9.self_attn.k_proj.bias": "0.safetensors",
262
+ "encoder.layers.9.self_attn.v_proj.weight": "0.safetensors",
263
+ "encoder.layers.9.self_attn.v_proj.scales": "0.safetensors",
264
+ "encoder.layers.9.self_attn.v_proj.biases": "0.safetensors",
265
+ "encoder.layers.9.self_attn.v_proj.bias": "0.safetensors",
266
+ "encoder.layers.9.self_attn.o_proj.weight": "0.safetensors",
267
+ "encoder.layers.9.self_attn.o_proj.scales": "0.safetensors",
268
+ "encoder.layers.9.self_attn.o_proj.biases": "0.safetensors",
269
+ "encoder.layers.9.self_attn.rotary_emb.inv_freq": "0.safetensors",
270
+ "encoder.layers.9.post_attention_layernorm.weight": "0.safetensors",
271
+ "encoder.layers.9.mlp.gate_proj.weight": "0.safetensors",
272
+ "encoder.layers.9.mlp.gate_proj.scales": "0.safetensors",
273
+ "encoder.layers.9.mlp.gate_proj.biases": "0.safetensors",
274
+ "encoder.layers.9.mlp.up_proj.weight": "0.safetensors",
275
+ "encoder.layers.9.mlp.up_proj.scales": "0.safetensors",
276
+ "encoder.layers.9.mlp.up_proj.biases": "0.safetensors",
277
+ "encoder.layers.9.mlp.down_proj.weight": "0.safetensors",
278
+ "encoder.layers.9.mlp.down_proj.scales": "0.safetensors",
279
+ "encoder.layers.9.mlp.down_proj.biases": "0.safetensors",
280
+ "encoder.layers.10.input_layernorm.weight": "0.safetensors",
281
+ "encoder.layers.10.self_attn.q_proj.weight": "0.safetensors",
282
+ "encoder.layers.10.self_attn.q_proj.scales": "0.safetensors",
283
+ "encoder.layers.10.self_attn.q_proj.biases": "0.safetensors",
284
+ "encoder.layers.10.self_attn.q_proj.bias": "0.safetensors",
285
+ "encoder.layers.10.self_attn.k_proj.weight": "0.safetensors",
286
+ "encoder.layers.10.self_attn.k_proj.scales": "0.safetensors",
287
+ "encoder.layers.10.self_attn.k_proj.biases": "0.safetensors",
288
+ "encoder.layers.10.self_attn.k_proj.bias": "0.safetensors",
289
+ "encoder.layers.10.self_attn.v_proj.weight": "0.safetensors",
290
+ "encoder.layers.10.self_attn.v_proj.scales": "0.safetensors",
291
+ "encoder.layers.10.self_attn.v_proj.biases": "0.safetensors",
292
+ "encoder.layers.10.self_attn.v_proj.bias": "0.safetensors",
293
+ "encoder.layers.10.self_attn.o_proj.weight": "0.safetensors",
294
+ "encoder.layers.10.self_attn.o_proj.scales": "0.safetensors",
295
+ "encoder.layers.10.self_attn.o_proj.biases": "0.safetensors",
296
+ "encoder.layers.10.self_attn.rotary_emb.inv_freq": "0.safetensors",
297
+ "encoder.layers.10.post_attention_layernorm.weight": "0.safetensors",
298
+ "encoder.layers.10.mlp.gate_proj.weight": "0.safetensors",
299
+ "encoder.layers.10.mlp.gate_proj.scales": "0.safetensors",
300
+ "encoder.layers.10.mlp.gate_proj.biases": "0.safetensors",
301
+ "encoder.layers.10.mlp.up_proj.weight": "0.safetensors",
302
+ "encoder.layers.10.mlp.up_proj.scales": "0.safetensors",
303
+ "encoder.layers.10.mlp.up_proj.biases": "0.safetensors",
304
+ "encoder.layers.10.mlp.down_proj.weight": "0.safetensors",
305
+ "encoder.layers.10.mlp.down_proj.scales": "0.safetensors",
306
+ "encoder.layers.10.mlp.down_proj.biases": "0.safetensors",
307
+ "encoder.layers.11.input_layernorm.weight": "0.safetensors",
308
+ "encoder.layers.11.self_attn.q_proj.weight": "0.safetensors",
309
+ "encoder.layers.11.self_attn.q_proj.scales": "0.safetensors",
310
+ "encoder.layers.11.self_attn.q_proj.biases": "0.safetensors",
311
+ "encoder.layers.11.self_attn.q_proj.bias": "0.safetensors",
312
+ "encoder.layers.11.self_attn.k_proj.weight": "0.safetensors",
313
+ "encoder.layers.11.self_attn.k_proj.scales": "0.safetensors",
314
+ "encoder.layers.11.self_attn.k_proj.biases": "0.safetensors",
315
+ "encoder.layers.11.self_attn.k_proj.bias": "0.safetensors",
316
+ "encoder.layers.11.self_attn.v_proj.weight": "0.safetensors",
317
+ "encoder.layers.11.self_attn.v_proj.scales": "0.safetensors",
318
+ "encoder.layers.11.self_attn.v_proj.biases": "0.safetensors",
319
+ "encoder.layers.11.self_attn.v_proj.bias": "0.safetensors",
320
+ "encoder.layers.11.self_attn.o_proj.weight": "0.safetensors",
321
+ "encoder.layers.11.self_attn.o_proj.scales": "0.safetensors",
322
+ "encoder.layers.11.self_attn.o_proj.biases": "0.safetensors",
323
+ "encoder.layers.11.self_attn.rotary_emb.inv_freq": "0.safetensors",
324
+ "encoder.layers.11.post_attention_layernorm.weight": "0.safetensors",
325
+ "encoder.layers.11.mlp.gate_proj.weight": "0.safetensors",
326
+ "encoder.layers.11.mlp.gate_proj.scales": "0.safetensors",
327
+ "encoder.layers.11.mlp.gate_proj.biases": "0.safetensors",
328
+ "encoder.layers.11.mlp.up_proj.weight": "0.safetensors",
329
+ "encoder.layers.11.mlp.up_proj.scales": "0.safetensors",
330
+ "encoder.layers.11.mlp.up_proj.biases": "0.safetensors",
331
+ "encoder.layers.11.mlp.down_proj.weight": "0.safetensors",
332
+ "encoder.layers.11.mlp.down_proj.scales": "0.safetensors",
333
+ "encoder.layers.11.mlp.down_proj.biases": "0.safetensors",
334
+ "encoder.layers.12.input_layernorm.weight": "0.safetensors",
335
+ "encoder.layers.12.self_attn.q_proj.weight": "0.safetensors",
336
+ "encoder.layers.12.self_attn.q_proj.scales": "0.safetensors",
337
+ "encoder.layers.12.self_attn.q_proj.biases": "0.safetensors",
338
+ "encoder.layers.12.self_attn.q_proj.bias": "0.safetensors",
339
+ "encoder.layers.12.self_attn.k_proj.weight": "0.safetensors",
340
+ "encoder.layers.12.self_attn.k_proj.scales": "0.safetensors",
341
+ "encoder.layers.12.self_attn.k_proj.biases": "0.safetensors",
342
+ "encoder.layers.12.self_attn.k_proj.bias": "0.safetensors",
343
+ "encoder.layers.12.self_attn.v_proj.weight": "0.safetensors",
344
+ "encoder.layers.12.self_attn.v_proj.scales": "0.safetensors",
345
+ "encoder.layers.12.self_attn.v_proj.biases": "0.safetensors",
346
+ "encoder.layers.12.self_attn.v_proj.bias": "0.safetensors",
347
+ "encoder.layers.12.self_attn.o_proj.weight": "0.safetensors",
348
+ "encoder.layers.12.self_attn.o_proj.scales": "0.safetensors",
349
+ "encoder.layers.12.self_attn.o_proj.biases": "0.safetensors",
350
+ "encoder.layers.12.self_attn.rotary_emb.inv_freq": "0.safetensors",
351
+ "encoder.layers.12.post_attention_layernorm.weight": "0.safetensors",
352
+ "encoder.layers.12.mlp.gate_proj.weight": "0.safetensors",
353
+ "encoder.layers.12.mlp.gate_proj.scales": "0.safetensors",
354
+ "encoder.layers.12.mlp.gate_proj.biases": "0.safetensors",
355
+ "encoder.layers.12.mlp.up_proj.weight": "0.safetensors",
356
+ "encoder.layers.12.mlp.up_proj.scales": "0.safetensors",
357
+ "encoder.layers.12.mlp.up_proj.biases": "0.safetensors",
358
+ "encoder.layers.12.mlp.down_proj.weight": "0.safetensors",
359
+ "encoder.layers.12.mlp.down_proj.scales": "0.safetensors",
360
+ "encoder.layers.12.mlp.down_proj.biases": "0.safetensors",
361
+ "encoder.layers.13.input_layernorm.weight": "0.safetensors",
362
+ "encoder.layers.13.self_attn.q_proj.weight": "0.safetensors",
363
+ "encoder.layers.13.self_attn.q_proj.scales": "0.safetensors",
364
+ "encoder.layers.13.self_attn.q_proj.biases": "0.safetensors",
365
+ "encoder.layers.13.self_attn.q_proj.bias": "0.safetensors",
366
+ "encoder.layers.13.self_attn.k_proj.weight": "0.safetensors",
367
+ "encoder.layers.13.self_attn.k_proj.scales": "0.safetensors",
368
+ "encoder.layers.13.self_attn.k_proj.biases": "0.safetensors",
369
+ "encoder.layers.13.self_attn.k_proj.bias": "0.safetensors",
370
+ "encoder.layers.13.self_attn.v_proj.weight": "0.safetensors",
371
+ "encoder.layers.13.self_attn.v_proj.scales": "0.safetensors",
372
+ "encoder.layers.13.self_attn.v_proj.biases": "0.safetensors",
373
+ "encoder.layers.13.self_attn.v_proj.bias": "0.safetensors",
374
+ "encoder.layers.13.self_attn.o_proj.weight": "0.safetensors",
375
+ "encoder.layers.13.self_attn.o_proj.scales": "0.safetensors",
376
+ "encoder.layers.13.self_attn.o_proj.biases": "0.safetensors",
377
+ "encoder.layers.13.self_attn.rotary_emb.inv_freq": "0.safetensors",
378
+ "encoder.layers.13.post_attention_layernorm.weight": "0.safetensors",
379
+ "encoder.layers.13.mlp.gate_proj.weight": "0.safetensors",
380
+ "encoder.layers.13.mlp.gate_proj.scales": "0.safetensors",
381
+ "encoder.layers.13.mlp.gate_proj.biases": "0.safetensors",
382
+ "encoder.layers.13.mlp.up_proj.weight": "0.safetensors",
383
+ "encoder.layers.13.mlp.up_proj.scales": "0.safetensors",
384
+ "encoder.layers.13.mlp.up_proj.biases": "0.safetensors",
385
+ "encoder.layers.13.mlp.down_proj.weight": "0.safetensors",
386
+ "encoder.layers.13.mlp.down_proj.scales": "0.safetensors",
387
+ "encoder.layers.13.mlp.down_proj.biases": "0.safetensors",
388
+ "encoder.layers.14.input_layernorm.weight": "0.safetensors",
389
+ "encoder.layers.14.self_attn.q_proj.weight": "1.safetensors",
390
+ "encoder.layers.14.self_attn.q_proj.scales": "1.safetensors",
391
+ "encoder.layers.14.self_attn.q_proj.biases": "1.safetensors",
392
+ "encoder.layers.14.self_attn.q_proj.bias": "1.safetensors",
393
+ "encoder.layers.14.self_attn.k_proj.weight": "1.safetensors",
394
+ "encoder.layers.14.self_attn.k_proj.scales": "1.safetensors",
395
+ "encoder.layers.14.self_attn.k_proj.biases": "1.safetensors",
396
+ "encoder.layers.14.self_attn.k_proj.bias": "1.safetensors",
397
+ "encoder.layers.14.self_attn.v_proj.weight": "1.safetensors",
398
+ "encoder.layers.14.self_attn.v_proj.scales": "1.safetensors",
399
+ "encoder.layers.14.self_attn.v_proj.biases": "1.safetensors",
400
+ "encoder.layers.14.self_attn.v_proj.bias": "1.safetensors",
401
+ "encoder.layers.14.self_attn.o_proj.weight": "1.safetensors",
402
+ "encoder.layers.14.self_attn.o_proj.scales": "1.safetensors",
403
+ "encoder.layers.14.self_attn.o_proj.biases": "1.safetensors",
404
+ "encoder.layers.14.self_attn.rotary_emb.inv_freq": "1.safetensors",
405
+ "encoder.layers.14.post_attention_layernorm.weight": "1.safetensors",
406
+ "encoder.layers.14.mlp.gate_proj.weight": "1.safetensors",
407
+ "encoder.layers.14.mlp.gate_proj.scales": "1.safetensors",
408
+ "encoder.layers.14.mlp.gate_proj.biases": "1.safetensors",
409
+ "encoder.layers.14.mlp.up_proj.weight": "1.safetensors",
410
+ "encoder.layers.14.mlp.up_proj.scales": "1.safetensors",
411
+ "encoder.layers.14.mlp.up_proj.biases": "1.safetensors",
412
+ "encoder.layers.14.mlp.down_proj.weight": "1.safetensors",
413
+ "encoder.layers.14.mlp.down_proj.scales": "1.safetensors",
414
+ "encoder.layers.14.mlp.down_proj.biases": "1.safetensors",
415
+ "encoder.layers.15.input_layernorm.weight": "1.safetensors",
416
+ "encoder.layers.15.self_attn.q_proj.weight": "1.safetensors",
417
+ "encoder.layers.15.self_attn.q_proj.scales": "1.safetensors",
418
+ "encoder.layers.15.self_attn.q_proj.biases": "1.safetensors",
419
+ "encoder.layers.15.self_attn.q_proj.bias": "1.safetensors",
420
+ "encoder.layers.15.self_attn.k_proj.weight": "1.safetensors",
421
+ "encoder.layers.15.self_attn.k_proj.scales": "1.safetensors",
422
+ "encoder.layers.15.self_attn.k_proj.biases": "1.safetensors",
423
+ "encoder.layers.15.self_attn.k_proj.bias": "1.safetensors",
424
+ "encoder.layers.15.self_attn.v_proj.weight": "1.safetensors",
425
+ "encoder.layers.15.self_attn.v_proj.scales": "1.safetensors",
426
+ "encoder.layers.15.self_attn.v_proj.biases": "1.safetensors",
427
+ "encoder.layers.15.self_attn.v_proj.bias": "1.safetensors",
428
+ "encoder.layers.15.self_attn.o_proj.weight": "1.safetensors",
429
+ "encoder.layers.15.self_attn.o_proj.scales": "1.safetensors",
430
+ "encoder.layers.15.self_attn.o_proj.biases": "1.safetensors",
431
+ "encoder.layers.15.self_attn.rotary_emb.inv_freq": "1.safetensors",
432
+ "encoder.layers.15.post_attention_layernorm.weight": "1.safetensors",
433
+ "encoder.layers.15.mlp.gate_proj.weight": "1.safetensors",
434
+ "encoder.layers.15.mlp.gate_proj.scales": "1.safetensors",
435
+ "encoder.layers.15.mlp.gate_proj.biases": "1.safetensors",
436
+ "encoder.layers.15.mlp.up_proj.weight": "1.safetensors",
437
+ "encoder.layers.15.mlp.up_proj.scales": "1.safetensors",
438
+ "encoder.layers.15.mlp.up_proj.biases": "1.safetensors",
439
+ "encoder.layers.15.mlp.down_proj.weight": "1.safetensors",
440
+ "encoder.layers.15.mlp.down_proj.scales": "1.safetensors",
441
+ "encoder.layers.15.mlp.down_proj.biases": "1.safetensors",
442
+ "encoder.layers.16.input_layernorm.weight": "1.safetensors",
443
+ "encoder.layers.16.self_attn.q_proj.weight": "1.safetensors",
444
+ "encoder.layers.16.self_attn.q_proj.scales": "1.safetensors",
445
+ "encoder.layers.16.self_attn.q_proj.biases": "1.safetensors",
446
+ "encoder.layers.16.self_attn.q_proj.bias": "1.safetensors",
447
+ "encoder.layers.16.self_attn.k_proj.weight": "1.safetensors",
448
+ "encoder.layers.16.self_attn.k_proj.scales": "1.safetensors",
449
+ "encoder.layers.16.self_attn.k_proj.biases": "1.safetensors",
450
+ "encoder.layers.16.self_attn.k_proj.bias": "1.safetensors",
451
+ "encoder.layers.16.self_attn.v_proj.weight": "1.safetensors",
452
+ "encoder.layers.16.self_attn.v_proj.scales": "1.safetensors",
453
+ "encoder.layers.16.self_attn.v_proj.biases": "1.safetensors",
454
+ "encoder.layers.16.self_attn.v_proj.bias": "1.safetensors",
455
+ "encoder.layers.16.self_attn.o_proj.weight": "1.safetensors",
456
+ "encoder.layers.16.self_attn.o_proj.scales": "1.safetensors",
457
+ "encoder.layers.16.self_attn.o_proj.biases": "1.safetensors",
458
+ "encoder.layers.16.self_attn.rotary_emb.inv_freq": "1.safetensors",
459
+ "encoder.layers.16.post_attention_layernorm.weight": "1.safetensors",
460
+ "encoder.layers.16.mlp.gate_proj.weight": "1.safetensors",
461
+ "encoder.layers.16.mlp.gate_proj.scales": "1.safetensors",
462
+ "encoder.layers.16.mlp.gate_proj.biases": "1.safetensors",
463
+ "encoder.layers.16.mlp.up_proj.weight": "1.safetensors",
464
+ "encoder.layers.16.mlp.up_proj.scales": "1.safetensors",
465
+ "encoder.layers.16.mlp.up_proj.biases": "1.safetensors",
466
+ "encoder.layers.16.mlp.down_proj.weight": "1.safetensors",
467
+ "encoder.layers.16.mlp.down_proj.scales": "1.safetensors",
468
+ "encoder.layers.16.mlp.down_proj.biases": "1.safetensors",
469
+ "encoder.layers.17.input_layernorm.weight": "1.safetensors",
470
+ "encoder.layers.17.self_attn.q_proj.weight": "1.safetensors",
471
+ "encoder.layers.17.self_attn.q_proj.scales": "1.safetensors",
472
+ "encoder.layers.17.self_attn.q_proj.biases": "1.safetensors",
473
+ "encoder.layers.17.self_attn.q_proj.bias": "1.safetensors",
474
+ "encoder.layers.17.self_attn.k_proj.weight": "1.safetensors",
475
+ "encoder.layers.17.self_attn.k_proj.scales": "1.safetensors",
476
+ "encoder.layers.17.self_attn.k_proj.biases": "1.safetensors",
477
+ "encoder.layers.17.self_attn.k_proj.bias": "1.safetensors",
478
+ "encoder.layers.17.self_attn.v_proj.weight": "1.safetensors",
479
+ "encoder.layers.17.self_attn.v_proj.scales": "1.safetensors",
480
+ "encoder.layers.17.self_attn.v_proj.biases": "1.safetensors",
481
+ "encoder.layers.17.self_attn.v_proj.bias": "1.safetensors",
482
+ "encoder.layers.17.self_attn.o_proj.weight": "1.safetensors",
483
+ "encoder.layers.17.self_attn.o_proj.scales": "1.safetensors",
484
+ "encoder.layers.17.self_attn.o_proj.biases": "1.safetensors",
485
+ "encoder.layers.17.self_attn.rotary_emb.inv_freq": "1.safetensors",
486
+ "encoder.layers.17.post_attention_layernorm.weight": "1.safetensors",
487
+ "encoder.layers.17.mlp.gate_proj.weight": "1.safetensors",
488
+ "encoder.layers.17.mlp.gate_proj.scales": "1.safetensors",
489
+ "encoder.layers.17.mlp.gate_proj.biases": "1.safetensors",
490
+ "encoder.layers.17.mlp.up_proj.weight": "1.safetensors",
491
+ "encoder.layers.17.mlp.up_proj.scales": "1.safetensors",
492
+ "encoder.layers.17.mlp.up_proj.biases": "1.safetensors",
493
+ "encoder.layers.17.mlp.down_proj.weight": "1.safetensors",
494
+ "encoder.layers.17.mlp.down_proj.scales": "1.safetensors",
495
+ "encoder.layers.17.mlp.down_proj.biases": "1.safetensors",
496
+ "encoder.layers.18.input_layernorm.weight": "1.safetensors",
497
+ "encoder.layers.18.self_attn.q_proj.weight": "1.safetensors",
498
+ "encoder.layers.18.self_attn.q_proj.scales": "1.safetensors",
499
+ "encoder.layers.18.self_attn.q_proj.biases": "1.safetensors",
500
+ "encoder.layers.18.self_attn.q_proj.bias": "1.safetensors",
501
+ "encoder.layers.18.self_attn.k_proj.weight": "1.safetensors",
502
+ "encoder.layers.18.self_attn.k_proj.scales": "1.safetensors",
503
+ "encoder.layers.18.self_attn.k_proj.biases": "1.safetensors",
504
+ "encoder.layers.18.self_attn.k_proj.bias": "1.safetensors",
505
+ "encoder.layers.18.self_attn.v_proj.weight": "1.safetensors",
506
+ "encoder.layers.18.self_attn.v_proj.scales": "1.safetensors",
507
+ "encoder.layers.18.self_attn.v_proj.biases": "1.safetensors",
508
+ "encoder.layers.18.self_attn.v_proj.bias": "1.safetensors",
509
+ "encoder.layers.18.self_attn.o_proj.weight": "1.safetensors",
510
+ "encoder.layers.18.self_attn.o_proj.scales": "1.safetensors",
511
+ "encoder.layers.18.self_attn.o_proj.biases": "1.safetensors",
512
+ "encoder.layers.18.self_attn.rotary_emb.inv_freq": "1.safetensors",
513
+ "encoder.layers.18.post_attention_layernorm.weight": "1.safetensors",
514
+ "encoder.layers.18.mlp.gate_proj.weight": "1.safetensors",
515
+ "encoder.layers.18.mlp.gate_proj.scales": "1.safetensors",
516
+ "encoder.layers.18.mlp.gate_proj.biases": "1.safetensors",
517
+ "encoder.layers.18.mlp.up_proj.weight": "1.safetensors",
518
+ "encoder.layers.18.mlp.up_proj.scales": "1.safetensors",
519
+ "encoder.layers.18.mlp.up_proj.biases": "1.safetensors",
520
+ "encoder.layers.18.mlp.down_proj.weight": "1.safetensors",
521
+ "encoder.layers.18.mlp.down_proj.scales": "1.safetensors",
522
+ "encoder.layers.18.mlp.down_proj.biases": "1.safetensors",
523
+ "encoder.layers.19.input_layernorm.weight": "1.safetensors",
524
+ "encoder.layers.19.self_attn.q_proj.weight": "1.safetensors",
525
+ "encoder.layers.19.self_attn.q_proj.scales": "1.safetensors",
526
+ "encoder.layers.19.self_attn.q_proj.biases": "1.safetensors",
527
+ "encoder.layers.19.self_attn.q_proj.bias": "1.safetensors",
528
+ "encoder.layers.19.self_attn.k_proj.weight": "1.safetensors",
529
+ "encoder.layers.19.self_attn.k_proj.scales": "1.safetensors",
530
+ "encoder.layers.19.self_attn.k_proj.biases": "1.safetensors",
531
+ "encoder.layers.19.self_attn.k_proj.bias": "1.safetensors",
532
+ "encoder.layers.19.self_attn.v_proj.weight": "1.safetensors",
533
+ "encoder.layers.19.self_attn.v_proj.scales": "1.safetensors",
534
+ "encoder.layers.19.self_attn.v_proj.biases": "1.safetensors",
535
+ "encoder.layers.19.self_attn.v_proj.bias": "1.safetensors",
536
+ "encoder.layers.19.self_attn.o_proj.weight": "1.safetensors",
537
+ "encoder.layers.19.self_attn.o_proj.scales": "1.safetensors",
538
+ "encoder.layers.19.self_attn.o_proj.biases": "1.safetensors",
539
+ "encoder.layers.19.self_attn.rotary_emb.inv_freq": "1.safetensors",
540
+ "encoder.layers.19.post_attention_layernorm.weight": "1.safetensors",
541
+ "encoder.layers.19.mlp.gate_proj.weight": "1.safetensors",
542
+ "encoder.layers.19.mlp.gate_proj.scales": "1.safetensors",
543
+ "encoder.layers.19.mlp.gate_proj.biases": "1.safetensors",
544
+ "encoder.layers.19.mlp.up_proj.weight": "1.safetensors",
545
+ "encoder.layers.19.mlp.up_proj.scales": "1.safetensors",
546
+ "encoder.layers.19.mlp.up_proj.biases": "1.safetensors",
547
+ "encoder.layers.19.mlp.down_proj.weight": "1.safetensors",
548
+ "encoder.layers.19.mlp.down_proj.scales": "1.safetensors",
549
+ "encoder.layers.19.mlp.down_proj.biases": "1.safetensors",
550
+ "encoder.layers.20.input_layernorm.weight": "1.safetensors",
551
+ "encoder.layers.20.self_attn.q_proj.weight": "1.safetensors",
552
+ "encoder.layers.20.self_attn.q_proj.scales": "1.safetensors",
553
+ "encoder.layers.20.self_attn.q_proj.biases": "1.safetensors",
554
+ "encoder.layers.20.self_attn.q_proj.bias": "1.safetensors",
555
+ "encoder.layers.20.self_attn.k_proj.weight": "1.safetensors",
556
+ "encoder.layers.20.self_attn.k_proj.scales": "1.safetensors",
557
+ "encoder.layers.20.self_attn.k_proj.biases": "1.safetensors",
558
+ "encoder.layers.20.self_attn.k_proj.bias": "1.safetensors",
559
+ "encoder.layers.20.self_attn.v_proj.weight": "1.safetensors",
560
+ "encoder.layers.20.self_attn.v_proj.scales": "1.safetensors",
561
+ "encoder.layers.20.self_attn.v_proj.biases": "1.safetensors",
562
+ "encoder.layers.20.self_attn.v_proj.bias": "1.safetensors",
563
+ "encoder.layers.20.self_attn.o_proj.weight": "1.safetensors",
564
+ "encoder.layers.20.self_attn.o_proj.scales": "1.safetensors",
565
+ "encoder.layers.20.self_attn.o_proj.biases": "1.safetensors",
566
+ "encoder.layers.20.self_attn.rotary_emb.inv_freq": "1.safetensors",
567
+ "encoder.layers.20.post_attention_layernorm.weight": "1.safetensors",
568
+ "encoder.layers.20.mlp.gate_proj.weight": "1.safetensors",
569
+ "encoder.layers.20.mlp.gate_proj.scales": "1.safetensors",
570
+ "encoder.layers.20.mlp.gate_proj.biases": "1.safetensors",
571
+ "encoder.layers.20.mlp.up_proj.weight": "1.safetensors",
572
+ "encoder.layers.20.mlp.up_proj.scales": "1.safetensors",
573
+ "encoder.layers.20.mlp.up_proj.biases": "1.safetensors",
574
+ "encoder.layers.20.mlp.down_proj.weight": "1.safetensors",
575
+ "encoder.layers.20.mlp.down_proj.scales": "1.safetensors",
576
+ "encoder.layers.20.mlp.down_proj.biases": "1.safetensors",
577
+ "encoder.layers.21.input_layernorm.weight": "1.safetensors",
578
+ "encoder.layers.21.self_attn.q_proj.weight": "1.safetensors",
579
+ "encoder.layers.21.self_attn.q_proj.scales": "1.safetensors",
580
+ "encoder.layers.21.self_attn.q_proj.biases": "1.safetensors",
581
+ "encoder.layers.21.self_attn.q_proj.bias": "1.safetensors",
582
+ "encoder.layers.21.self_attn.k_proj.weight": "1.safetensors",
583
+ "encoder.layers.21.self_attn.k_proj.scales": "1.safetensors",
584
+ "encoder.layers.21.self_attn.k_proj.biases": "1.safetensors",
585
+ "encoder.layers.21.self_attn.k_proj.bias": "1.safetensors",
586
+ "encoder.layers.21.self_attn.v_proj.weight": "1.safetensors",
587
+ "encoder.layers.21.self_attn.v_proj.scales": "1.safetensors",
588
+ "encoder.layers.21.self_attn.v_proj.biases": "1.safetensors",
589
+ "encoder.layers.21.self_attn.v_proj.bias": "1.safetensors",
590
+ "encoder.layers.21.self_attn.o_proj.weight": "1.safetensors",
591
+ "encoder.layers.21.self_attn.o_proj.scales": "1.safetensors",
592
+ "encoder.layers.21.self_attn.o_proj.biases": "1.safetensors",
593
+ "encoder.layers.21.self_attn.rotary_emb.inv_freq": "1.safetensors",
594
+ "encoder.layers.21.post_attention_layernorm.weight": "1.safetensors",
595
+ "encoder.layers.21.mlp.gate_proj.weight": "1.safetensors",
596
+ "encoder.layers.21.mlp.gate_proj.scales": "1.safetensors",
597
+ "encoder.layers.21.mlp.gate_proj.biases": "1.safetensors",
598
+ "encoder.layers.21.mlp.up_proj.weight": "1.safetensors",
599
+ "encoder.layers.21.mlp.up_proj.scales": "1.safetensors",
600
+ "encoder.layers.21.mlp.up_proj.biases": "1.safetensors",
601
+ "encoder.layers.21.mlp.down_proj.weight": "1.safetensors",
602
+ "encoder.layers.21.mlp.down_proj.scales": "1.safetensors",
603
+ "encoder.layers.21.mlp.down_proj.biases": "1.safetensors",
604
+ "encoder.layers.22.input_layernorm.weight": "1.safetensors",
605
+ "encoder.layers.22.self_attn.q_proj.weight": "1.safetensors",
606
+ "encoder.layers.22.self_attn.q_proj.scales": "1.safetensors",
607
+ "encoder.layers.22.self_attn.q_proj.biases": "1.safetensors",
608
+ "encoder.layers.22.self_attn.q_proj.bias": "1.safetensors",
609
+ "encoder.layers.22.self_attn.k_proj.weight": "1.safetensors",
610
+ "encoder.layers.22.self_attn.k_proj.scales": "1.safetensors",
611
+ "encoder.layers.22.self_attn.k_proj.biases": "1.safetensors",
612
+ "encoder.layers.22.self_attn.k_proj.bias": "1.safetensors",
613
+ "encoder.layers.22.self_attn.v_proj.weight": "1.safetensors",
614
+ "encoder.layers.22.self_attn.v_proj.scales": "1.safetensors",
615
+ "encoder.layers.22.self_attn.v_proj.biases": "1.safetensors",
616
+ "encoder.layers.22.self_attn.v_proj.bias": "1.safetensors",
617
+ "encoder.layers.22.self_attn.o_proj.weight": "1.safetensors",
618
+ "encoder.layers.22.self_attn.o_proj.scales": "1.safetensors",
619
+ "encoder.layers.22.self_attn.o_proj.biases": "1.safetensors",
620
+ "encoder.layers.22.self_attn.rotary_emb.inv_freq": "1.safetensors",
621
+ "encoder.layers.22.post_attention_layernorm.weight": "1.safetensors",
622
+ "encoder.layers.22.mlp.gate_proj.weight": "1.safetensors",
623
+ "encoder.layers.22.mlp.gate_proj.scales": "1.safetensors",
624
+ "encoder.layers.22.mlp.gate_proj.biases": "1.safetensors",
625
+ "encoder.layers.22.mlp.up_proj.weight": "1.safetensors",
626
+ "encoder.layers.22.mlp.up_proj.scales": "1.safetensors",
627
+ "encoder.layers.22.mlp.up_proj.biases": "1.safetensors",
628
+ "encoder.layers.22.mlp.down_proj.weight": "1.safetensors",
629
+ "encoder.layers.22.mlp.down_proj.scales": "1.safetensors",
630
+ "encoder.layers.22.mlp.down_proj.biases": "1.safetensors",
631
+ "encoder.layers.23.input_layernorm.weight": "1.safetensors",
632
+ "encoder.layers.23.self_attn.q_proj.weight": "1.safetensors",
633
+ "encoder.layers.23.self_attn.q_proj.scales": "1.safetensors",
634
+ "encoder.layers.23.self_attn.q_proj.biases": "1.safetensors",
635
+ "encoder.layers.23.self_attn.q_proj.bias": "1.safetensors",
636
+ "encoder.layers.23.self_attn.k_proj.weight": "1.safetensors",
637
+ "encoder.layers.23.self_attn.k_proj.scales": "1.safetensors",
638
+ "encoder.layers.23.self_attn.k_proj.biases": "1.safetensors",
639
+ "encoder.layers.23.self_attn.k_proj.bias": "1.safetensors",
640
+ "encoder.layers.23.self_attn.v_proj.weight": "1.safetensors",
641
+ "encoder.layers.23.self_attn.v_proj.scales": "1.safetensors",
642
+ "encoder.layers.23.self_attn.v_proj.biases": "1.safetensors",
643
+ "encoder.layers.23.self_attn.v_proj.bias": "1.safetensors",
644
+ "encoder.layers.23.self_attn.o_proj.weight": "1.safetensors",
645
+ "encoder.layers.23.self_attn.o_proj.scales": "1.safetensors",
646
+ "encoder.layers.23.self_attn.o_proj.biases": "1.safetensors",
647
+ "encoder.layers.23.self_attn.rotary_emb.inv_freq": "1.safetensors",
648
+ "encoder.layers.23.post_attention_layernorm.weight": "1.safetensors",
649
+ "encoder.layers.23.mlp.gate_proj.weight": "1.safetensors",
650
+ "encoder.layers.23.mlp.gate_proj.scales": "1.safetensors",
651
+ "encoder.layers.23.mlp.gate_proj.biases": "1.safetensors",
652
+ "encoder.layers.23.mlp.up_proj.weight": "1.safetensors",
653
+ "encoder.layers.23.mlp.up_proj.scales": "1.safetensors",
654
+ "encoder.layers.23.mlp.up_proj.biases": "1.safetensors",
655
+ "encoder.layers.23.mlp.down_proj.weight": "1.safetensors",
656
+ "encoder.layers.23.mlp.down_proj.scales": "1.safetensors",
657
+ "encoder.layers.23.mlp.down_proj.biases": "1.safetensors",
658
+ "encoder.layers.24.input_layernorm.weight": "1.safetensors",
659
+ "encoder.layers.24.self_attn.q_proj.weight": "1.safetensors",
660
+ "encoder.layers.24.self_attn.q_proj.scales": "1.safetensors",
661
+ "encoder.layers.24.self_attn.q_proj.biases": "1.safetensors",
662
+ "encoder.layers.24.self_attn.q_proj.bias": "1.safetensors",
663
+ "encoder.layers.24.self_attn.k_proj.weight": "1.safetensors",
664
+ "encoder.layers.24.self_attn.k_proj.scales": "1.safetensors",
665
+ "encoder.layers.24.self_attn.k_proj.biases": "1.safetensors",
666
+ "encoder.layers.24.self_attn.k_proj.bias": "1.safetensors",
667
+ "encoder.layers.24.self_attn.v_proj.weight": "1.safetensors",
668
+ "encoder.layers.24.self_attn.v_proj.scales": "1.safetensors",
669
+ "encoder.layers.24.self_attn.v_proj.biases": "1.safetensors",
670
+ "encoder.layers.24.self_attn.v_proj.bias": "1.safetensors",
671
+ "encoder.layers.24.self_attn.o_proj.weight": "1.safetensors",
672
+ "encoder.layers.24.self_attn.o_proj.scales": "1.safetensors",
673
+ "encoder.layers.24.self_attn.o_proj.biases": "1.safetensors",
674
+ "encoder.layers.24.self_attn.rotary_emb.inv_freq": "1.safetensors",
675
+ "encoder.layers.24.post_attention_layernorm.weight": "1.safetensors",
676
+ "encoder.layers.24.mlp.gate_proj.weight": "1.safetensors",
677
+ "encoder.layers.24.mlp.gate_proj.scales": "1.safetensors",
678
+ "encoder.layers.24.mlp.gate_proj.biases": "1.safetensors",
679
+ "encoder.layers.24.mlp.up_proj.weight": "1.safetensors",
680
+ "encoder.layers.24.mlp.up_proj.scales": "1.safetensors",
681
+ "encoder.layers.24.mlp.up_proj.biases": "1.safetensors",
682
+ "encoder.layers.24.mlp.down_proj.weight": "1.safetensors",
683
+ "encoder.layers.24.mlp.down_proj.scales": "1.safetensors",
684
+ "encoder.layers.24.mlp.down_proj.biases": "1.safetensors",
685
+ "encoder.layers.25.input_layernorm.weight": "1.safetensors",
686
+ "encoder.layers.25.self_attn.q_proj.weight": "1.safetensors",
687
+ "encoder.layers.25.self_attn.q_proj.scales": "1.safetensors",
688
+ "encoder.layers.25.self_attn.q_proj.biases": "1.safetensors",
689
+ "encoder.layers.25.self_attn.q_proj.bias": "1.safetensors",
690
+ "encoder.layers.25.self_attn.k_proj.weight": "1.safetensors",
691
+ "encoder.layers.25.self_attn.k_proj.scales": "1.safetensors",
692
+ "encoder.layers.25.self_attn.k_proj.biases": "1.safetensors",
693
+ "encoder.layers.25.self_attn.k_proj.bias": "1.safetensors",
694
+ "encoder.layers.25.self_attn.v_proj.weight": "1.safetensors",
695
+ "encoder.layers.25.self_attn.v_proj.scales": "1.safetensors",
696
+ "encoder.layers.25.self_attn.v_proj.biases": "1.safetensors",
697
+ "encoder.layers.25.self_attn.v_proj.bias": "1.safetensors",
698
+ "encoder.layers.25.self_attn.o_proj.weight": "1.safetensors",
699
+ "encoder.layers.25.self_attn.o_proj.scales": "1.safetensors",
700
+ "encoder.layers.25.self_attn.o_proj.biases": "1.safetensors",
701
+ "encoder.layers.25.self_attn.rotary_emb.inv_freq": "1.safetensors",
702
+ "encoder.layers.25.post_attention_layernorm.weight": "1.safetensors",
703
+ "encoder.layers.25.mlp.gate_proj.weight": "1.safetensors",
704
+ "encoder.layers.25.mlp.gate_proj.scales": "1.safetensors",
705
+ "encoder.layers.25.mlp.gate_proj.biases": "1.safetensors",
706
+ "encoder.layers.25.mlp.up_proj.weight": "1.safetensors",
707
+ "encoder.layers.25.mlp.up_proj.scales": "1.safetensors",
708
+ "encoder.layers.25.mlp.up_proj.biases": "1.safetensors",
709
+ "encoder.layers.25.mlp.down_proj.weight": "1.safetensors",
710
+ "encoder.layers.25.mlp.down_proj.scales": "1.safetensors",
711
+ "encoder.layers.25.mlp.down_proj.biases": "1.safetensors",
712
+ "encoder.layers.26.input_layernorm.weight": "1.safetensors",
713
+ "encoder.layers.26.self_attn.q_proj.weight": "1.safetensors",
714
+ "encoder.layers.26.self_attn.q_proj.scales": "1.safetensors",
715
+ "encoder.layers.26.self_attn.q_proj.biases": "1.safetensors",
716
+ "encoder.layers.26.self_attn.q_proj.bias": "1.safetensors",
717
+ "encoder.layers.26.self_attn.k_proj.weight": "1.safetensors",
718
+ "encoder.layers.26.self_attn.k_proj.scales": "1.safetensors",
719
+ "encoder.layers.26.self_attn.k_proj.biases": "1.safetensors",
720
+ "encoder.layers.26.self_attn.k_proj.bias": "1.safetensors",
721
+ "encoder.layers.26.self_attn.v_proj.weight": "1.safetensors",
722
+ "encoder.layers.26.self_attn.v_proj.scales": "1.safetensors",
723
+ "encoder.layers.26.self_attn.v_proj.biases": "1.safetensors",
724
+ "encoder.layers.26.self_attn.v_proj.bias": "1.safetensors",
725
+ "encoder.layers.26.self_attn.o_proj.weight": "1.safetensors",
726
+ "encoder.layers.26.self_attn.o_proj.scales": "1.safetensors",
727
+ "encoder.layers.26.self_attn.o_proj.biases": "1.safetensors",
728
+ "encoder.layers.26.self_attn.rotary_emb.inv_freq": "1.safetensors",
729
+ "encoder.layers.26.post_attention_layernorm.weight": "1.safetensors",
730
+ "encoder.layers.26.mlp.gate_proj.weight": "1.safetensors",
731
+ "encoder.layers.26.mlp.gate_proj.scales": "1.safetensors",
732
+ "encoder.layers.26.mlp.gate_proj.biases": "1.safetensors",
733
+ "encoder.layers.26.mlp.up_proj.weight": "1.safetensors",
734
+ "encoder.layers.26.mlp.up_proj.scales": "1.safetensors",
735
+ "encoder.layers.26.mlp.up_proj.biases": "1.safetensors",
736
+ "encoder.layers.26.mlp.down_proj.weight": "1.safetensors",
737
+ "encoder.layers.26.mlp.down_proj.scales": "1.safetensors",
738
+ "encoder.layers.26.mlp.down_proj.biases": "1.safetensors",
739
+ "encoder.layers.27.input_layernorm.weight": "1.safetensors",
740
+ "encoder.layers.27.self_attn.q_proj.weight": "1.safetensors",
741
+ "encoder.layers.27.self_attn.q_proj.scales": "1.safetensors",
742
+ "encoder.layers.27.self_attn.q_proj.biases": "1.safetensors",
743
+ "encoder.layers.27.self_attn.q_proj.bias": "1.safetensors",
744
+ "encoder.layers.27.self_attn.k_proj.weight": "1.safetensors",
745
+ "encoder.layers.27.self_attn.k_proj.scales": "1.safetensors",
746
+ "encoder.layers.27.self_attn.k_proj.biases": "1.safetensors",
747
+ "encoder.layers.27.self_attn.k_proj.bias": "1.safetensors",
748
+ "encoder.layers.27.self_attn.v_proj.weight": "1.safetensors",
749
+ "encoder.layers.27.self_attn.v_proj.scales": "1.safetensors",
750
+ "encoder.layers.27.self_attn.v_proj.biases": "1.safetensors",
751
+ "encoder.layers.27.self_attn.v_proj.bias": "1.safetensors",
752
+ "encoder.layers.27.self_attn.o_proj.weight": "1.safetensors",
753
+ "encoder.layers.27.self_attn.o_proj.scales": "1.safetensors",
754
+ "encoder.layers.27.self_attn.o_proj.biases": "1.safetensors",
755
+ "encoder.layers.27.self_attn.rotary_emb.inv_freq": "1.safetensors",
756
+ "encoder.layers.27.post_attention_layernorm.weight": "1.safetensors",
757
+ "encoder.layers.27.mlp.gate_proj.weight": "1.safetensors",
758
+ "encoder.layers.27.mlp.gate_proj.scales": "1.safetensors",
759
+ "encoder.layers.27.mlp.gate_proj.biases": "1.safetensors",
760
+ "encoder.layers.27.mlp.up_proj.weight": "1.safetensors",
761
+ "encoder.layers.27.mlp.up_proj.scales": "1.safetensors",
762
+ "encoder.layers.27.mlp.up_proj.biases": "1.safetensors",
763
+ "encoder.layers.27.mlp.down_proj.weight": "1.safetensors",
764
+ "encoder.layers.27.mlp.down_proj.scales": "1.safetensors",
765
+ "encoder.layers.27.mlp.down_proj.biases": "1.safetensors",
766
+ "encoder.norm.weight": "1.safetensors",
767
+ "encoder.rotary_emb.inv_freq": "1.safetensors"
768
+ }
769
+ }
tokenizer/tokenizer.json ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:444a80c37c7f34f485c050cf374595a0cb7d88735459cf0505fd204453c33178
3
+ size 11417006
tokenizer/tokenizer_config.json ADDED
@@ -0,0 +1,10 @@
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "add_prefix_space": null,
3
+ "backend": "tokenizers",
4
+ "bos_token": null,
5
+ "eos_token": "<|endoftext|>",
6
+ "model_max_length": 1000000000000000019884624838656,
7
+ "pad_token": "<|endoftext|>",
8
+ "tokenizer_class": "Qwen2Tokenizer",
9
+ "unk_token": "<|endoftext|>"
10
+ }
transformer/0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b8c616f786d2516c1c01b0a05c816cac89c8dcde355311f46cb038f81b449979
3
+ size 2133769869
transformer/1.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:0a42229f780dd021c1a7879778416caf08c538acdaeaa9918084fa12453dbdf2
3
+ size 2142940100
transformer/2.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:d380a48d79cab74b19638665e828e634eb64fa00f0dc0eb9a96acaecbe9702bd
3
+ size 2103978106
transformer/3.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:6345e1f12aa8039ac033e6d8ba1f54dc4bcfe20753abfc4678db2c7162a8d6a0
3
+ size 2121682265
transformer/4.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:a31365721624ab1ed5b750dcbaf31266bd45818fbc80faae09c8cfe0968a5385
3
+ size 2142940118
transformer/5.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:b1c77b152696e99c1e9e18a35d40140a5d9c9df76dd62c905d368513ea71157a
3
+ size 2103978210
transformer/6.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:03f189f37e4fa16d1e4e59f50225ac1912e980c0ba2e0a38312ec81751bedc50
3
+ size 449941880
transformer/model.safetensors.index.json ADDED
The diff for this file is too large to render. See raw diff
 
vae/0.safetensors ADDED
@@ -0,0 +1,3 @@
 
 
 
 
1
+ version https://git-lfs.github.com/spec/v1
2
+ oid sha256:2c2db147e652a0e011c1c5f40345ab84e9a95819ee9c7b29eb29f245cccd8f64
3
+ size 253586147
vae/model.safetensors.index.json ADDED
@@ -0,0 +1,200 @@
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
1
+ {
2
+ "metadata": {
3
+ "quantization_level": "4",
4
+ "mflux_version": "0.18.2"
5
+ },
6
+ "weight_map": {
7
+ "decoder.conv_in.conv3d.weight": "0.safetensors",
8
+ "decoder.conv_in.conv3d.bias": "0.safetensors",
9
+ "decoder.mid_block.attentions.0.norm.weight": "0.safetensors",
10
+ "decoder.mid_block.attentions.0.to_qkv.weight": "0.safetensors",
11
+ "decoder.mid_block.attentions.0.to_qkv.bias": "0.safetensors",
12
+ "decoder.mid_block.attentions.0.proj.weight": "0.safetensors",
13
+ "decoder.mid_block.attentions.0.proj.bias": "0.safetensors",
14
+ "decoder.mid_block.resnets.0.norm1.weight": "0.safetensors",
15
+ "decoder.mid_block.resnets.0.conv1.conv3d.weight": "0.safetensors",
16
+ "decoder.mid_block.resnets.0.conv1.conv3d.bias": "0.safetensors",
17
+ "decoder.mid_block.resnets.0.norm2.weight": "0.safetensors",
18
+ "decoder.mid_block.resnets.0.conv2.conv3d.weight": "0.safetensors",
19
+ "decoder.mid_block.resnets.0.conv2.conv3d.bias": "0.safetensors",
20
+ "decoder.mid_block.resnets.1.norm1.weight": "0.safetensors",
21
+ "decoder.mid_block.resnets.1.conv1.conv3d.weight": "0.safetensors",
22
+ "decoder.mid_block.resnets.1.conv1.conv3d.bias": "0.safetensors",
23
+ "decoder.mid_block.resnets.1.norm2.weight": "0.safetensors",
24
+ "decoder.mid_block.resnets.1.conv2.conv3d.weight": "0.safetensors",
25
+ "decoder.mid_block.resnets.1.conv2.conv3d.bias": "0.safetensors",
26
+ "decoder.up_block0.resnets.0.norm1.weight": "0.safetensors",
27
+ "decoder.up_block0.resnets.0.conv1.conv3d.weight": "0.safetensors",
28
+ "decoder.up_block0.resnets.0.conv1.conv3d.bias": "0.safetensors",
29
+ "decoder.up_block0.resnets.0.norm2.weight": "0.safetensors",
30
+ "decoder.up_block0.resnets.0.conv2.conv3d.weight": "0.safetensors",
31
+ "decoder.up_block0.resnets.0.conv2.conv3d.bias": "0.safetensors",
32
+ "decoder.up_block0.resnets.1.norm1.weight": "0.safetensors",
33
+ "decoder.up_block0.resnets.1.conv1.conv3d.weight": "0.safetensors",
34
+ "decoder.up_block0.resnets.1.conv1.conv3d.bias": "0.safetensors",
35
+ "decoder.up_block0.resnets.1.norm2.weight": "0.safetensors",
36
+ "decoder.up_block0.resnets.1.conv2.conv3d.weight": "0.safetensors",
37
+ "decoder.up_block0.resnets.1.conv2.conv3d.bias": "0.safetensors",
38
+ "decoder.up_block0.resnets.2.norm1.weight": "0.safetensors",
39
+ "decoder.up_block0.resnets.2.conv1.conv3d.weight": "0.safetensors",
40
+ "decoder.up_block0.resnets.2.conv1.conv3d.bias": "0.safetensors",
41
+ "decoder.up_block0.resnets.2.norm2.weight": "0.safetensors",
42
+ "decoder.up_block0.resnets.2.conv2.conv3d.weight": "0.safetensors",
43
+ "decoder.up_block0.resnets.2.conv2.conv3d.bias": "0.safetensors",
44
+ "decoder.up_block0.upsamplers.0.time_conv.conv3d.weight": "0.safetensors",
45
+ "decoder.up_block0.upsamplers.0.time_conv.conv3d.bias": "0.safetensors",
46
+ "decoder.up_block0.upsamplers.0.resample_conv.weight": "0.safetensors",
47
+ "decoder.up_block0.upsamplers.0.resample_conv.bias": "0.safetensors",
48
+ "decoder.up_block1.resnets.0.norm1.weight": "0.safetensors",
49
+ "decoder.up_block1.resnets.0.conv1.conv3d.weight": "0.safetensors",
50
+ "decoder.up_block1.resnets.0.conv1.conv3d.bias": "0.safetensors",
51
+ "decoder.up_block1.resnets.0.norm2.weight": "0.safetensors",
52
+ "decoder.up_block1.resnets.0.conv2.conv3d.weight": "0.safetensors",
53
+ "decoder.up_block1.resnets.0.conv2.conv3d.bias": "0.safetensors",
54
+ "decoder.up_block1.resnets.0.skip_conv.conv3d.weight": "0.safetensors",
55
+ "decoder.up_block1.resnets.0.skip_conv.conv3d.bias": "0.safetensors",
56
+ "decoder.up_block1.resnets.1.norm1.weight": "0.safetensors",
57
+ "decoder.up_block1.resnets.1.conv1.conv3d.weight": "0.safetensors",
58
+ "decoder.up_block1.resnets.1.conv1.conv3d.bias": "0.safetensors",
59
+ "decoder.up_block1.resnets.1.norm2.weight": "0.safetensors",
60
+ "decoder.up_block1.resnets.1.conv2.conv3d.weight": "0.safetensors",
61
+ "decoder.up_block1.resnets.1.conv2.conv3d.bias": "0.safetensors",
62
+ "decoder.up_block1.resnets.2.norm1.weight": "0.safetensors",
63
+ "decoder.up_block1.resnets.2.conv1.conv3d.weight": "0.safetensors",
64
+ "decoder.up_block1.resnets.2.conv1.conv3d.bias": "0.safetensors",
65
+ "decoder.up_block1.resnets.2.norm2.weight": "0.safetensors",
66
+ "decoder.up_block1.resnets.2.conv2.conv3d.weight": "0.safetensors",
67
+ "decoder.up_block1.resnets.2.conv2.conv3d.bias": "0.safetensors",
68
+ "decoder.up_block1.upsamplers.0.time_conv.conv3d.weight": "0.safetensors",
69
+ "decoder.up_block1.upsamplers.0.time_conv.conv3d.bias": "0.safetensors",
70
+ "decoder.up_block1.upsamplers.0.resample_conv.weight": "0.safetensors",
71
+ "decoder.up_block1.upsamplers.0.resample_conv.bias": "0.safetensors",
72
+ "decoder.up_block2.resnets.0.norm1.weight": "0.safetensors",
73
+ "decoder.up_block2.resnets.0.conv1.conv3d.weight": "0.safetensors",
74
+ "decoder.up_block2.resnets.0.conv1.conv3d.bias": "0.safetensors",
75
+ "decoder.up_block2.resnets.0.norm2.weight": "0.safetensors",
76
+ "decoder.up_block2.resnets.0.conv2.conv3d.weight": "0.safetensors",
77
+ "decoder.up_block2.resnets.0.conv2.conv3d.bias": "0.safetensors",
78
+ "decoder.up_block2.resnets.1.norm1.weight": "0.safetensors",
79
+ "decoder.up_block2.resnets.1.conv1.conv3d.weight": "0.safetensors",
80
+ "decoder.up_block2.resnets.1.conv1.conv3d.bias": "0.safetensors",
81
+ "decoder.up_block2.resnets.1.norm2.weight": "0.safetensors",
82
+ "decoder.up_block2.resnets.1.conv2.conv3d.weight": "0.safetensors",
83
+ "decoder.up_block2.resnets.1.conv2.conv3d.bias": "0.safetensors",
84
+ "decoder.up_block2.resnets.2.norm1.weight": "0.safetensors",
85
+ "decoder.up_block2.resnets.2.conv1.conv3d.weight": "0.safetensors",
86
+ "decoder.up_block2.resnets.2.conv1.conv3d.bias": "0.safetensors",
87
+ "decoder.up_block2.resnets.2.norm2.weight": "0.safetensors",
88
+ "decoder.up_block2.resnets.2.conv2.conv3d.weight": "0.safetensors",
89
+ "decoder.up_block2.resnets.2.conv2.conv3d.bias": "0.safetensors",
90
+ "decoder.up_block2.upsamplers.0.resample_conv.weight": "0.safetensors",
91
+ "decoder.up_block2.upsamplers.0.resample_conv.bias": "0.safetensors",
92
+ "decoder.up_block3.resnets.0.norm1.weight": "0.safetensors",
93
+ "decoder.up_block3.resnets.0.conv1.conv3d.weight": "0.safetensors",
94
+ "decoder.up_block3.resnets.0.conv1.conv3d.bias": "0.safetensors",
95
+ "decoder.up_block3.resnets.0.norm2.weight": "0.safetensors",
96
+ "decoder.up_block3.resnets.0.conv2.conv3d.weight": "0.safetensors",
97
+ "decoder.up_block3.resnets.0.conv2.conv3d.bias": "0.safetensors",
98
+ "decoder.up_block3.resnets.1.norm1.weight": "0.safetensors",
99
+ "decoder.up_block3.resnets.1.conv1.conv3d.weight": "0.safetensors",
100
+ "decoder.up_block3.resnets.1.conv1.conv3d.bias": "0.safetensors",
101
+ "decoder.up_block3.resnets.1.norm2.weight": "0.safetensors",
102
+ "decoder.up_block3.resnets.1.conv2.conv3d.weight": "0.safetensors",
103
+ "decoder.up_block3.resnets.1.conv2.conv3d.bias": "0.safetensors",
104
+ "decoder.up_block3.resnets.2.norm1.weight": "0.safetensors",
105
+ "decoder.up_block3.resnets.2.conv1.conv3d.weight": "0.safetensors",
106
+ "decoder.up_block3.resnets.2.conv1.conv3d.bias": "0.safetensors",
107
+ "decoder.up_block3.resnets.2.norm2.weight": "0.safetensors",
108
+ "decoder.up_block3.resnets.2.conv2.conv3d.weight": "0.safetensors",
109
+ "decoder.up_block3.resnets.2.conv2.conv3d.bias": "0.safetensors",
110
+ "decoder.norm_out.weight": "0.safetensors",
111
+ "decoder.conv_out.conv3d.weight": "0.safetensors",
112
+ "decoder.conv_out.conv3d.bias": "0.safetensors",
113
+ "encoder.conv_in.conv3d.weight": "0.safetensors",
114
+ "encoder.conv_in.conv3d.bias": "0.safetensors",
115
+ "encoder.down_blocks.0.resnets.0.norm1.weight": "0.safetensors",
116
+ "encoder.down_blocks.0.resnets.0.conv1.conv3d.weight": "0.safetensors",
117
+ "encoder.down_blocks.0.resnets.0.conv1.conv3d.bias": "0.safetensors",
118
+ "encoder.down_blocks.0.resnets.0.norm2.weight": "0.safetensors",
119
+ "encoder.down_blocks.0.resnets.0.conv2.conv3d.weight": "0.safetensors",
120
+ "encoder.down_blocks.0.resnets.0.conv2.conv3d.bias": "0.safetensors",
121
+ "encoder.down_blocks.0.resnets.1.norm1.weight": "0.safetensors",
122
+ "encoder.down_blocks.0.resnets.1.conv1.conv3d.weight": "0.safetensors",
123
+ "encoder.down_blocks.0.resnets.1.conv1.conv3d.bias": "0.safetensors",
124
+ "encoder.down_blocks.0.resnets.1.norm2.weight": "0.safetensors",
125
+ "encoder.down_blocks.0.resnets.1.conv2.conv3d.weight": "0.safetensors",
126
+ "encoder.down_blocks.0.resnets.1.conv2.conv3d.bias": "0.safetensors",
127
+ "encoder.down_blocks.0.downsamplers.0.resample_conv.weight": "0.safetensors",
128
+ "encoder.down_blocks.0.downsamplers.0.resample_conv.bias": "0.safetensors",
129
+ "encoder.down_blocks.1.resnets.0.norm1.weight": "0.safetensors",
130
+ "encoder.down_blocks.1.resnets.0.conv1.conv3d.weight": "0.safetensors",
131
+ "encoder.down_blocks.1.resnets.0.conv1.conv3d.bias": "0.safetensors",
132
+ "encoder.down_blocks.1.resnets.0.norm2.weight": "0.safetensors",
133
+ "encoder.down_blocks.1.resnets.0.conv2.conv3d.weight": "0.safetensors",
134
+ "encoder.down_blocks.1.resnets.0.conv2.conv3d.bias": "0.safetensors",
135
+ "encoder.down_blocks.1.resnets.0.skip_conv.conv3d.weight": "0.safetensors",
136
+ "encoder.down_blocks.1.resnets.0.skip_conv.conv3d.bias": "0.safetensors",
137
+ "encoder.down_blocks.1.resnets.1.norm1.weight": "0.safetensors",
138
+ "encoder.down_blocks.1.resnets.1.conv1.conv3d.weight": "0.safetensors",
139
+ "encoder.down_blocks.1.resnets.1.conv1.conv3d.bias": "0.safetensors",
140
+ "encoder.down_blocks.1.resnets.1.norm2.weight": "0.safetensors",
141
+ "encoder.down_blocks.1.resnets.1.conv2.conv3d.weight": "0.safetensors",
142
+ "encoder.down_blocks.1.resnets.1.conv2.conv3d.bias": "0.safetensors",
143
+ "encoder.down_blocks.1.downsamplers.0.resample_conv.weight": "0.safetensors",
144
+ "encoder.down_blocks.1.downsamplers.0.resample_conv.bias": "0.safetensors",
145
+ "encoder.down_blocks.2.resnets.0.norm1.weight": "0.safetensors",
146
+ "encoder.down_blocks.2.resnets.0.conv1.conv3d.weight": "0.safetensors",
147
+ "encoder.down_blocks.2.resnets.0.conv1.conv3d.bias": "0.safetensors",
148
+ "encoder.down_blocks.2.resnets.0.norm2.weight": "0.safetensors",
149
+ "encoder.down_blocks.2.resnets.0.conv2.conv3d.weight": "0.safetensors",
150
+ "encoder.down_blocks.2.resnets.0.conv2.conv3d.bias": "0.safetensors",
151
+ "encoder.down_blocks.2.resnets.0.skip_conv.conv3d.weight": "0.safetensors",
152
+ "encoder.down_blocks.2.resnets.0.skip_conv.conv3d.bias": "0.safetensors",
153
+ "encoder.down_blocks.2.resnets.1.norm1.weight": "0.safetensors",
154
+ "encoder.down_blocks.2.resnets.1.conv1.conv3d.weight": "0.safetensors",
155
+ "encoder.down_blocks.2.resnets.1.conv1.conv3d.bias": "0.safetensors",
156
+ "encoder.down_blocks.2.resnets.1.norm2.weight": "0.safetensors",
157
+ "encoder.down_blocks.2.resnets.1.conv2.conv3d.weight": "0.safetensors",
158
+ "encoder.down_blocks.2.resnets.1.conv2.conv3d.bias": "0.safetensors",
159
+ "encoder.down_blocks.2.downsamplers.0.time_conv.conv3d.weight": "0.safetensors",
160
+ "encoder.down_blocks.2.downsamplers.0.time_conv.conv3d.bias": "0.safetensors",
161
+ "encoder.down_blocks.2.downsamplers.0.resample_conv.weight": "0.safetensors",
162
+ "encoder.down_blocks.2.downsamplers.0.resample_conv.bias": "0.safetensors",
163
+ "encoder.down_blocks.3.resnets.0.norm1.weight": "0.safetensors",
164
+ "encoder.down_blocks.3.resnets.0.conv1.conv3d.weight": "0.safetensors",
165
+ "encoder.down_blocks.3.resnets.0.conv1.conv3d.bias": "0.safetensors",
166
+ "encoder.down_blocks.3.resnets.0.norm2.weight": "0.safetensors",
167
+ "encoder.down_blocks.3.resnets.0.conv2.conv3d.weight": "0.safetensors",
168
+ "encoder.down_blocks.3.resnets.0.conv2.conv3d.bias": "0.safetensors",
169
+ "encoder.down_blocks.3.resnets.1.norm1.weight": "0.safetensors",
170
+ "encoder.down_blocks.3.resnets.1.conv1.conv3d.weight": "0.safetensors",
171
+ "encoder.down_blocks.3.resnets.1.conv1.conv3d.bias": "0.safetensors",
172
+ "encoder.down_blocks.3.resnets.1.norm2.weight": "0.safetensors",
173
+ "encoder.down_blocks.3.resnets.1.conv2.conv3d.weight": "0.safetensors",
174
+ "encoder.down_blocks.3.resnets.1.conv2.conv3d.bias": "0.safetensors",
175
+ "encoder.mid_block.attentions.0.norm.weight": "0.safetensors",
176
+ "encoder.mid_block.attentions.0.to_qkv.weight": "0.safetensors",
177
+ "encoder.mid_block.attentions.0.to_qkv.bias": "0.safetensors",
178
+ "encoder.mid_block.attentions.0.proj.weight": "0.safetensors",
179
+ "encoder.mid_block.attentions.0.proj.bias": "0.safetensors",
180
+ "encoder.mid_block.resnets.0.norm1.weight": "0.safetensors",
181
+ "encoder.mid_block.resnets.0.conv1.conv3d.weight": "0.safetensors",
182
+ "encoder.mid_block.resnets.0.conv1.conv3d.bias": "0.safetensors",
183
+ "encoder.mid_block.resnets.0.norm2.weight": "0.safetensors",
184
+ "encoder.mid_block.resnets.0.conv2.conv3d.weight": "0.safetensors",
185
+ "encoder.mid_block.resnets.0.conv2.conv3d.bias": "0.safetensors",
186
+ "encoder.mid_block.resnets.1.norm1.weight": "0.safetensors",
187
+ "encoder.mid_block.resnets.1.conv1.conv3d.weight": "0.safetensors",
188
+ "encoder.mid_block.resnets.1.conv1.conv3d.bias": "0.safetensors",
189
+ "encoder.mid_block.resnets.1.norm2.weight": "0.safetensors",
190
+ "encoder.mid_block.resnets.1.conv2.conv3d.weight": "0.safetensors",
191
+ "encoder.mid_block.resnets.1.conv2.conv3d.bias": "0.safetensors",
192
+ "encoder.norm_out.weight": "0.safetensors",
193
+ "encoder.conv_out.conv3d.weight": "0.safetensors",
194
+ "encoder.conv_out.conv3d.bias": "0.safetensors",
195
+ "post_quant_conv.conv3d.weight": "0.safetensors",
196
+ "post_quant_conv.conv3d.bias": "0.safetensors",
197
+ "quant_conv.conv3d.weight": "0.safetensors",
198
+ "quant_conv.conv3d.bias": "0.safetensors"
199
+ }
200
+ }